Skip to main content

Usability testing guide

How to test your product with real users

Usability testing reveals how real users interact with your product. Learn how to plan, run, and analyze usability tests — even with zero budget.

What is usability testing?

Usability testing means observing real users attempting to complete tasks with your product. The goal is to find where users get confused, stuck, or frustrated — before your product launches. You are not asking for opinions. You are watching behavior.

Why 5 users is enough

Research by Jakob Nielsen shows that 5 users typically uncover 85% of usability problems. Testing with more people finds diminishing returns — the same problems surface repeatedly while new findings become rare.

The right cadence: test with 5, fix what you find, then test with another 5. Iterative small rounds beat one large study almost every time.

Types of usability testing

Each approach has different tradeoffs between depth, speed, and logistics. Most teams mix moderated and unmoderated depending on the stage of the product.

Moderated

Early-stage

You (the researcher) guide the session, ask follow-up questions, and probe for reasoning. Best for early-stage and complex flows where you need to understand the 'why' behind behavior — not just what happened.

Unmoderated

Scalable

Users complete tasks independently using a tool like Maze, UserTesting, or Lyssna. Faster and easier to run at scale, but you lose the ability to ask follow-up questions or probe for reasoning.

Remote

Easier to recruit

Remote sessions are easier to recruit for and require no physical space. The tradeoff: you cannot observe body language, and screen sharing introduces occasional technical friction.

In-person

Highest depth

In-person testing reveals body language and subtle friction that remote sessions miss — sighs, grimaces, moments of hesitation before a click. Worth the logistics for high-stakes flows.

How to run a moderated session

A moderated usability session follows five steps. The most important discipline: stay silent during tasks, even when you want to help.

Step 1

Recruit

Find 5 users who match your target persona. LinkedIn, Prolific, or ask your existing customers. You do not need a huge panel — you need the right people.

Step 2

Write scenarios

Task descriptions that feel natural: 'Imagine you just signed up. Find out how to change your password.' Never say what to click. The moment you point to a button, you have destroyed the test.

Step 3

Run the session

Introduce yourself, explain the think-aloud protocol ('Please say out loud what you are thinking and feeling'), then observe silently. You are there to watch, not guide.

Step 4

Take notes

Log what users say, do, and where they pause or backtrack. Do not help them — watching them struggle is the data. Every hesitation is a finding.

Step 5

Debrief

Ask 2–3 open questions after the session: 'What was the hardest part?' 'What would you change?' This surfaces problems they did not verbalize during the task.

The think-aloud protocol

Ask participants to verbalize their thoughts as they work: "I am looking for a settings button... I do not see it... maybe it is behind my profile picture?"

This narration reveals mental models and friction points that silent observation misses. When a user says "I expected this to be here" — that is a design finding. When they say nothing and pause for ten seconds — that is also a finding. The protocol surfaces both.

Analyzing results

Raw session notes are not findings. Analysis is the step that turns observations into prioritized problems your team can act on.

1

Create an affinity map

Group similar observations on sticky notes (Miro or FigJam). Cluster them into themes: navigation confusion, label clarity, missing feedback, etc.

2

Prioritize by frequency and severity

Frequency: how many users hit this problem? Severity: did it block them completely, or just slow them down? Problems that are both frequent and severe go to the top of the fix list.

3

Report format

For each finding: Problem → User quote → Recommendation. The quote grounds the finding in observed evidence — not your opinion about the design.

Free usability testing tools

You do not need a budget to run usability testing. These tools cover both moderated and unmoderated sessions at no cost.

MazeFree plan — unmoderated testing with click maps and task flows
LyssnaFree plan — first-click testing, preference tests, surveys
Google Meet + screen shareFree — use for moderated remote sessions
LookbackSession recording with highlight clipping for stakeholder sharing

Apply research in practice

Apply research in the UX Designer track

Usability testing is a core skill for UX designers. The designer track covers research methods, information architecture, prototyping, and design systems — end to end.

UX / UI Designer track